New opportunity

Agentic AI Delivery Leader

Ascensus · Atlanta, United States

About this role

The Agentic AI Delivery Leader is a Senior Manager in Ascensus’ Transformation team, responsible for delivering an enterprise portfolio of AI agent initiatives from discovery and business case through design, deployment, adoption, and measured outcomes. The role works with business executives, product and engineering teams, data and AI practitioners, and risk and compliance partners to improve customer experience, associate productivity, and operational performance.

Responsibilities

Own a multi-quarter portfolio roadmap, identify high-value opportunities, define success metrics, and balance business value, quality, scalability, and delivery trade-offs. Assess user journeys and data and system constraints; frame solutions using agent tools and actions, guardrails, and autonomous or human-in-the-loop approaches.
Develop funding proposals and business cases covering costs, revenue, and customer experience. Establish baselines and attribution for benefits measurement, then track outcomes through adoption and continuous improvement.
Coordinate product, AI engineering, data science, architecture, operations, and external partners to deliver pilots and scaled deployments. Manage scope, dependencies, risks, quality, executive steering, and program governance using Agile or hybrid delivery methods.
Work with the Responsible AI leader on data handling, model and vendor risk, explainability, resilience, content filtering, and escalation controls. Lead launch planning, training, communications, hypercare, stakeholder readiness, and post-launch feedback and enhancements.
Advise executives on operating-model changes and investment decisions; lead governance forums, manage cross-functional dependencies, and guide build, buy, or partner decisions and vendor performance. The role supervises an Agentic AI Delivery Principal and is responsible for protecting confidential data.

Requirements

A bachelor’s degree and 8+ years in digital transformation, product or program management, or consulting are required, along with 4+ years delivering AI/ML and generative AI solutions in production and 1+ year delivering agentic AI solutions in production. Candidates need experience with LLM-powered agents or copilots across discovery, evaluation, deployment, and iteration; understanding of RAG, MCP/API tool calling, autonomous AI, human-in-the-loop processes, and prompt/configuration management; and the ability to lead concurrent workstreams, executive communications, delivery metrics, testing, and release readiness. Strong analytical, communication, stakeholder-influence, and coaching skills are required. An MBA or equivalent advanced degree is preferred.

This is a full-time, hybrid position in Atlanta, Georgia. Ascensus does not provide visa sponsorship and does not hire candidates working under OPT.

For full details about this role, click “Apply on company site” to view the employer’s original posting.

Skills for this role

Agentic AIAI agentsGenerative AIMachine LearningLarge Language ModelsRAGMCPAPIsTool callingFunction callingHuman-in-the-loop systemsPrompt managementAI evaluationResponsible AIAgile deliveryProgram managementPortfolio managementProduct managementDigital transformationBusiness case developmentBenefits realizationStakeholder managementExecutive communicationChange managementVendor managementData governanceRisk managementExperimentation

Checking your profile…

Preparation suggestions, not the employer’s actual interview questions. Always check the original posting for current requirements.

Understand the role

What matters most

Portfolio strategy and investment
Own a multi-quarter roadmap of agentic AI use cases, prioritize for business value and scalability, and defend funding cases covering cost, revenue, and customer experience.
Practical agent solution design
Use user journeys and data and system constraints to frame solutions, including tool actions, guardrails, and choices between autonomy and human review.
Cross-functional delivery governance
Coordinate product, AI engineering, data, architecture, operations, and partners from pilots to scaled deployments; manage dependencies, quality, executive steering, and delivery commitments.
Responsible production deployment
Work with the Responsible AI leader on data handling, model and vendor risk, explainability, resilience, filtering, and escalation. Protect confidential data.
Adoption and measured outcomes
Lead launch readiness, training, communications, and hypercare; establish benefit baselines and attribution, track adoption, and use feedback to guide improvements.

Likely challenges to think through

Choosing the right level of agent autonomy
A likely trade-off is speed and usefulness versus accuracy, data protection, and escalation risk. Prepare to explain how you would evaluate a use case, set acceptance criteria, and decide where human review is needed.
Proving value beyond a successful pilot
Benefits may be hard to attribute if workflows and adoption change alongside the technology. Prepare an approach to baselines, measurement, post-launch feedback, and decisions about whether to scale or revise a deployment.
Keeping concurrent initiatives aligned
Competing priorities and dependencies across business, technical, risk, and vendor teams could put scope or timelines under pressure. Prepare an example of how you would make trade-offs visible to executives and secure decisions through governance.

Practice your answers

Technical practice

How would you prioritize a multi-quarter portfolio of agentic AI use cases?

What to coverSuggested preparation: compare business value, quality, scalability, portability, dependencies, and time to value. Explain how you would set success metrics and revisit trade-offs as evidence changes.

How would you discover and frame an opportunity for an AI agent?

What to coverSuggested preparation: walk through user journeys, operational pain points, diagnostics, and data or system constraints before proposing a solution. Explain how you would define the agent’s tools, actions, guardrails, and delivery plan.

How would you choose and evaluate RAG, MCP/API tool calling, and prompt configuration for an LLM-powered agent?

What to coverSuggested preparation: distinguish retrieving information from taking actions, and discuss how source quality, tool permissions, and configuration changes affect reliability. Describe how you would test the chosen approach against the use case rather than assume every capability is needed.

Where would you draw the line between autonomous actions and human-in-the-loop review?

What to coverSuggested preparation: weigh the consequences of an incorrect action, the need for timely service, and whether a reviewer can make a meaningful decision. Explain escalation and guardrail choices for higher-risk cases.

How would you evaluate an agent from pilot through production iteration?

What to coverSuggested preparation: cover acceptance criteria, testing of accuracy and quality, failure cases, and release-readiness decisions. Explain how production feedback and measured outcomes would inform prompt or configuration changes and the enhancement backlog.

How would you build a funding case and measure whether an agent delivered its promised benefits?

What to coverSuggested preparation: account for costs and potential revenue, customer-experience, and productivity effects without assuming all benefits are immediately measurable. Describe baselines, attribution, adoption measures, and how you would address outcomes that differ from the forecast.

How would you govern concurrent pilots and scaled deployments across several teams?

What to coverSuggested preparation: describe an Agile or hybrid cadence that makes scope, dependencies, risks, quality, and decisions visible to delivery teams and executive steering. Explain what you would escalate and how you would keep commitments realistic.

What controls would you require before an agent handles confidential data in production?

What to coverSuggested preparation: discuss data handling, prevention of unauthorized access or disclosure, content filtering, explainability, resilience, and escalation paths. Explain how you would work with the Responsible AI leader on model and vendor risk rather than treat delivery and controls as separate work.

How would you plan go-live and determine whether an agent is being adopted?

What to coverSuggested preparation: address stakeholder readiness, training, communications, hypercare, and feedback collection. Distinguish a successful technical release from sustained use and measured business outcomes.

How would you make a build, buy, or partner decision for an agent capability?

What to coverSuggested preparation: compare delivery speed, fit, scalability, risk, and ongoing performance against the use case and portfolio roadmap. Explain how contract scope and vendor performance would be managed against delivery outcomes.

Behavioural practice

Tell me about a time you secured executive alignment on an uncertain investment.

What to coverSuggested preparation: use situation, action, and result to show how you presented a value case, made trade-offs clear, and obtained or revised a decision. Focus on what you communicated and what the decision enabled.

Tell me about a time you resolved a dependency across teams you did not directly manage.

What to coverSuggested preparation: describe the situation, your actions with the affected teams, and the result for scope or delivery. Show how you influenced without authority while keeping competing priorities visible.

Tell me about a time a delivery or quality risk forced you to change course.

What to coverSuggested preparation: explain the situation, how you assessed and escalated the risk, the action you took, and the result. Address the trade-off among timeline, quality, and business commitment.

Tell me about a time you led adoption after launching a new capability.

What to coverSuggested preparation: give a situation, the training, communications, or feedback actions you led, and a measurable result where available. Explain what you changed when initial use or stakeholder readiness fell short.

Tell me about a time you coached a colleague while leading complex delivery.

What to coverSuggested preparation: describe the situation, specific coaching or delegation actions, and the result for the colleague and the work. Keep the example relevant to the role’s supervision and cross-functional leadership responsibilities.

Sharpen your knowledge

Topics to prepare

Agent architecture and integration
Revise how RAG, prompt/configuration management, and MCP/API tool or function calling fit together in an LLM-powered agent. As a revision exercise, sketch a hypothetical agent flow and explain what each component contributes. The posting explicitly calls for understanding these elements and experience taking agents through deployment and iteration.
Autonomy and human oversight
Compare autonomous actions with human-in-the-loop approval for a hypothetical workflow. Identify where guardrails, escalation paths, and human review would sit. This reflects the role’s responsibility for framing agent solutions around tools, actions, and appropriate oversight.
Portfolio strategy and governance
Review how a multi-quarter roadmap can balance business value, quality, scalability, and portability. Practice explaining why two hypothetical initiatives would be sequenced differently. The role owns an enterprise portfolio and its governance, metrics, and delivery trade-offs.
Benefits measurement and attribution
Work through a hypothetical baseline, success metric, and attribution question for customer experience or productivity. Explain how adoption and experimentation would affect the interpretation of results. The posting requires business cases, benefits tracking, and measured outcomes.
Responsible AI in production
Review the distinctions among data handling, model and vendor risk, explainability, resilience, content filtering, and escalation controls. For revision, identify a plausible failure mode and the control that would address it. The role partners with the Responsible AI leader and must protect confidential data.

Skills to revise

Frame an agent opportunity
Practice turning a hypothetical user journey and pain point into a one-page discovery brief covering data and system constraints, proposed agent actions, guardrails, and a delivery plan. This mirrors the posting’s opportunity-discovery and solution-framing duties.
Build an executive value case
Draft a short hypothetical funding proposal showing costs, potential revenue or customer-experience impact, a baseline, and how benefits would be measured. The role must build and defend business cases and secure executive alignment.
Run a concurrent delivery plan
Create a sample multi-quarter roadmap with two workstreams, dependencies, risks, decision points, and an Agile or hybrid operating cadence; summarize the trade-offs in an executive steering update. The posting emphasizes cross-functional delivery and program governance.
Define evaluation and release criteria
Write sample acceptance criteria and a testing plan for a hypothetical LLM-powered agent, including quality, accuracy, escalation behavior, and release readiness. The requirements call for agent evaluation and direct work with product, engineering, and data teams on testing and releases.
Plan launch and iteration
Draft a launch checklist covering stakeholder readiness, training, communications, hypercare, adoption measures, and a post-launch feedback backlog. These are stated go-live and change-leadership responsibilities.
Present a build, buy, or partner decision
Prepare a concise comparison of hypothetical options using delivery outcomes, contract scope, vendor performance, and risk as decision factors. The posting explicitly assigns vendor and platform decisions to this role.

Your preparation checklist

Map the stated requirements to evidence
Create a one-page reference to examples you could discuss for production AI/ML, generative AI and agentic AI delivery, concurrent workstreams, executive communication and coaching. The posting specifies 8+ years in transformation, product/program management or consulting; 4+ years delivering AI/ML and generative AI in production; and 1+ year delivering agentic AI in production.
Prepare an end-to-end agent delivery walkthrough
Select a relevant project and outline discovery, evaluation, deployment, adoption and post-launch iteration. Bring the decisions you made, your role, delivery metrics and measured outcomes; omit confidential information.
Practice framing an agent opportunity
For a hypothetical customer or associate journey, draft a one-page design showing the pain point, data and system constraints, RAG or tool-calling choices, agent actions, guardrails and where human review belongs. Present it as an interview exercise, not an Ascensus project.
Draft a portfolio and value-case exercise
Sketch a multi-quarter prioritization approach and one illustrative funding case covering costs, potential revenue or customer-experience effects, baseline, attribution and success metrics. Identify the trade-offs you would take to an executive decision forum.
Prepare a production-risk discussion
Outline how you would partner with a Responsible AI leader on confidential-data handling, model and vendor risk, explainability, resilience, content filtering and escalation. Include evaluation, testing and release-readiness criteria for your example agent.
Rehearse an executive steering update
Make a short sample update for concurrent pilots and scaled deployments: decisions needed, scope, dependencies, risks, delivery metrics and build/buy/partner considerations. Be ready to explain how you would coordinate squads and coach the supervised Agentic AI Delivery Principal.
Outline launch and benefits tracking
Prepare a sample go-live plan with stakeholder readiness, training, communications, hypercare, adoption measures and a post-launch feedback backlog. Show how results would be compared with the value-case baseline.
Write targeted questions for Ascensus
Ask how the team prioritizes agent opportunities, defines production quality and portability, shares decisions with Responsible AI partners, and measures adoption and benefits across business units. Keep these as questions rather than assumptions about current practices.

Availability last checked: Nov 4, 2026. Status can change after a check; confirm on the employer’s website. Report an issue · Our checking policy

KEEP EXPLORING

Similar AI jobs

Related skills and specializations in United States. Matched to this role, not your profile.

Explore more jobs

Finding similar opportunities…